Multi-modal fusion intelligent diagnosis and prediction method for faults of speed reducer of strip steel rolling mill

Abstract

The invention relates to the technical field of industrial equipment fault diagnosis, in particular to a multi-modal fusion intelligent diagnosis and prediction method for faults of a speed reducer of a strip steel rolling mill. Comprising the following steps: giving a unique identifier to a speed reducer of a strip mill, obtaining sensor and image multi-modal data, and establishing a speed reducer data mapping table; constructing a diagnosis model, obtaining a speed reducer data mapping table, and outputting speed reducer fault diagnosis and prediction data; acquiring speed reducer associated data, and dynamically outputting the working mode of the strip mill; according to the fault diagnosis and prediction of the speed reducer of the strip mill, the working mode of the strip mill is dynamically adjusted, the non-planned shutdown condition is reduced, and the strip steel production rhythm is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment fault diagnosis, and particularly to an intelligent fault diagnosis and prediction method for a reducer of a strip rolling mill with multi-modal fusion. Background Art

[0002] The working characteristics of a special reducer for a rolling mill are low speed, heavy load, large impact load, frequent impact times, and continuous operation. In modern industrial production, as a key equipment, the operating state of the reducer of a strip rolling mill directly affects production efficiency and product quality. However, during operation, the reducer is easily affected by various factors such as mechanical wear and electrical interference, resulting in frequent failures. Traditional fault diagnosis methods rely on manual experience and single signal sources, which are difficult to comprehensively and accurately reflect the actual operating state of the reducer, and the diagnostic results have great uncertainty.

[0003] Currently, relevant data of the reducer can be obtained by sensors. For example, vibration signals of the reducer can be collected by vibration sensors, temperature data can be collected by temperature sensors, and oil condition information can be collected by oil sensors. However, after obtaining the relevant data, the current and subsequent working modes of the strip rolling mill still need to be adjusted manually, resulting in more unplanned shutdowns and poor strip production rhythm. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent fault diagnosis and prediction method for a reducer of a strip rolling mill with multi-modal fusion, aiming to solve the technical problem that after obtaining the relevant data in the prior art, the current and subsequent working modes of the strip rolling mill still need to be adjusted manually, resulting in more unplanned shutdowns and poor strip production rhythm.

[0005] To achieve the above purpose, an intelligent fault diagnosis and prediction method for a reducer of a strip rolling mill with multi-modal fusion adopted by the present invention includes the following steps: Assign a unique identifier to the reducer of the strip rolling mill, obtain multi-modal data of sensors and images, and establish a data mapping table for the reducer; Construct a diagnostic model, obtain the data mapping table of the reducer, and output fault diagnosis and prediction data of the reducer; Obtain associated data of the reducer and dynamically output the working mode of the strip rolling mill.

[0006] Among them, in the step of assigning a unique identifier to the reducer of the strip rolling mill, obtaining multi-modal data of sensors and images, and establishing a data mapping table for the reducer: Define an encoding rule and assign a unique identifier to the reducer according to the information of the strip rolling mill; A variety of sensing devices and image acquisition devices are set at the reducer to respectively obtain the temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the reducer. A reducer data mapping table is established with the unique identifier as the primary key.

[0007] Among them, in the step of setting a variety of sensing devices and image acquisition devices at the reducer to respectively obtain the temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the reducer: Time stamps and the unique identifier of the reducer are assigned to the data obtained by the sensing devices and image acquisition devices.

[0008] Among them, before the step of establishing a reducer data mapping table with the unique identifier as the primary key: Using the unique identifier as the main search term, query the data of the unique identifier and perform data aggregation.

[0009] Among them, in the step of constructing a diagnostic model, obtaining the reducer data mapping table, and outputting the reducer fault diagnosis and prediction data: Obtain the historical data of the reducer and construct a diagnostic model using the historical data of the reducer; Using the diagnostic model, according to the reducer data mapping table, obtain the reducer fault diagnosis data and output it; Using the diagnostic model, according to the reducer data mapping table, obtain the reducer fault prediction data and output it.

[0010] Among them, in the step of obtaining the associated data of the reducer and dynamically outputting the working mode of the strip mill: According to the cooperation data of multiple reducers in the steel rolling mill, each reducer in a single and multiple steel rolling mills is associated; Divide the task types of the steel rolling mill and define the priorities of the task types; Divide the operation area, according to the reducer fault diagnosis and prediction data, obtain the emergency task type, query the current task instructions of multiple steel rolling mills in the operation area, compare the priorities of the current task type and the emergency task type, and output the next operation instruction for the steel rolling mill.

[0011] Among them, in the step of dividing the task types of the steel rolling mill and defining the priorities of the task types: The task types include emergency tasks, timing tasks, and temporary tasks; among them, the priority of the emergency task is level one, the priority of the timing task is level two, and the priority of the temporary task is level three.

[0012] Among them, in the step of comparing the priorities of the current task type and the emergency task type and outputting the next operation instruction for the steel rolling mill: When the priority of the current task type is higher than or equal to that of the emergency task type, an instruction to continue the current task is output for the steel rolling mill.

[0013] Among them, in the step of comparing the priorities of the current task type and the emergency task type and outputting the next operation instruction for the steel rolling mill: When the priority of the current task type is lower than that of the emergency task type, an instruction to terminate the current task is output for the steel rolling mill, and the current task data is packaged.

[0014] A method for intelligent diagnosis and prediction of faults in a strip steel rolling mill reducer with multi-modal fusion according to the present invention assigns a unique identifier to the strip steel rolling mill reducer, obtains multi-modal data of sensors and images, and establishes a data mapping table for the reducer; constructs a diagnostic model, obtains the data mapping table for the reducer, and outputs fault diagnosis and prediction data for the reducer; obtains associated data of the reducer and dynamically outputs the working mode of the strip steel rolling mill; by dynamically adjusting the working mode of the strip steel rolling mill according to the fault diagnosis and prediction of the strip steel rolling mill reducer, the occurrence of unplanned shutdowns is reduced, and the strip steel production rhythm is optimized. Brief Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a flowchart of the steps of the method for intelligent diagnosis and prediction of faults in a strip steel rolling mill reducer with multi-modal fusion according to the present invention.

[0017] Figure 2 It is a flowchart of the steps of S100 of the present invention.

[0018] Figure 3 It is a flowchart of the steps of S200 of the present invention.

[0019] Figure 4 It is a flowchart of the steps of S300 of the present invention.

[0020] Figure 5 It is a schematic structural diagram of the system for intelligent diagnosis and prediction of faults in a strip steel rolling mill reducer with multi-modal fusion according to the present invention.

[0021] Figure 6 It is a schematic structural diagram of the electronic device according to the present invention.

[0022] 401 - Data mapping table establishment module, 402 - Diagnosis and prediction module, 403 - Mode dynamic output module. Detailed implementation manners

[0023] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application.

[0024] The terms used in the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a", "the", and "said" used in the present application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0025] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0026] Please refer to Figures 1 to 4 , the present invention provides a method for intelligent fault diagnosis and prediction of a strip mill reducer with multimodal fusion, including the following steps: S100: Assign a unique identifier to the strip mill reducer, obtain multi-modal data of sensors and images, and establish a reducer data mapping table.

[0027] In this embodiment, a unique identifier is assigned to the strip mill reducer, multi-modal data of sensors and images are obtained, and a reducer data mapping table is established. The specific process is as follows: S101: Define a coding rule, and assign a unique identifier to the reducer according to the information of the strip mill; S102: Set a variety of sensing devices and image acquisition devices at the reducer, respectively obtain the temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the reducer, and assign timestamps and the unique identifier of the reducer to the data obtained by the sensing devices and image acquisition devices; S103: Use the unique identifier as the main search term, query the data of the unique identifier, and perform data aggregation; S104: Use the unique identifier as the primary key to establish a reducer data mapping table.

[0028] During the above process, define the coding rules and assign a unique identifier to the speed reducer according to the information of the strip mill. In defining the coding rules: it is necessary to avoid duplicate identifiers, and at the same time, the identifier should contain key information such as equipment type, location, production batch, etc., which is convenient for manual identification. For example: Identifier structure: Equipment type - Mill number - Speed reducer number - Production batch Equipment type: Represented by letters. For example, D represents the speed reducer; Mill number: Represented by numbers. For example, 01 represents the first mill; Speed reducer number: Represented by numbers. For example, 03 represents the third speed reducer; Production batch: Represented by year and month. For example, 2501 represents production in January 2025; Example: D - 01 - 03 - 2501.

[0029] Set a variety of sensing devices and image acquisition devices at the speed reducer. Set a temperature sensor, a pressure sensor, a vibration amplitude sensor, an oil level sensor, and a camera at the speed reducer to obtain the temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the speed reducer respectively. Among them: The temperature sensor is used for: monitoring the temperature of the speed reducer to prevent overheating; The pressure sensor is used for: monitoring the pressure of the lubricating oil to ensure the normal operation of the lubrication system; The vibration amplitude sensor is used for: monitoring the vibration of the speed reducer to give early warning of mechanical failures; The oil level sensor is used for: monitoring the lubricating oil level to prevent insufficient oil volume; The camera is used for: collecting lubricating oil images and analyzing the change of oil quality.

[0030] All sensors and cameras collect data in real time and transmit it to the data acquisition system through an industrial bus (such as Modbus, OPC UA).

[0031] Add a timestamp to each piece of data to record the collection time, and at the same time add the unique identifier of the speed reducer to each piece of data to ensure data traceability.

[0032] Among them, the data format example: Temperature data: {"id":"D - 01 - 03 - 2501","timestamp":"2025 - 01 - 15T12:34:56","temperature": 75.2}.

[0033] Image data: {"id":"D-01-03-2501","timestamp":"2025-01-15T12:34:56","image_url":"http: / / example.com / image123.jpg"}。

[0034] Use the unique identifier as the main search term to query the data of this unique identifier and perform data aggregation; among them, data query includes database query and real-time stream processing: Database query: Retrieve all historical data of this reducer from the database through the unique identifier, such as D-01-03-2501.

[0035] Real-time stream processing: For real-time data such as vibration data, perform real-time queries through stream processing frameworks such as Kafka and Flink.

[0036] During the data aggregation process, perform correlation analysis on multi-source data of temperature, pressure, vibration, oil level, and image, and extract key features from the data, such as vibration amplitude trend, oil level change law, etc.

[0037] Use the unique identifier as the primary key to establish a reducer data mapping table. The structure of the data mapping table is as follows: Primary key: Unique identifier, such as D-01-03-2501; Field design: id: Unique identifier; timestamp: Data acquisition time; temperature: Temperature data; pressure: Pressure data; vibration: Vibration data; oil_level: Oil level data; image_url: Image data storage path.

[0038] The reducer data mapping table adopts a dynamic update mechanism: Insert real-time data, and realize real-time data writing through a message queue such as RabbitMQ.

[0039] Regular full update, set the update time, such as updating historical data every 6h, 12h, or 24h.

[0040] Mark abnormal data, and specially mark the data that exceeds the threshold, such as vibration exceeding the standard.

[0041] In this step, a unique identifier is generated for each speed reducer in the strip mill, and real-time data is collected through multi-modal sensors. Finally, a data mapping table with the unique identifier as the primary key is established to achieve the full life cycle management of the speed reducer status.

[0042] S200: Build a diagnostic model, obtain the data mapping table of the speed reducer, and output the speed reducer fault diagnosis and prediction data.

[0043] In this embodiment, a diagnostic model is built, the data mapping table of the speed reducer is obtained, and the speed reducer fault diagnosis and prediction data are output. The specific process is as follows: S201: Obtain the historical data of the speed reducer and build a diagnostic model using the historical data of the speed reducer; S202: Use the diagnostic model to obtain the speed reducer fault diagnosis data according to the data mapping table of the speed reducer and output it; S203: Use the diagnostic model to obtain the speed reducer fault prediction data according to the data mapping table of the speed reducer and output it.

[0044] In the above process, determine the source of the historical data of the speed reducer. For example, obtain the historical data of the speed reducer from the sensor data acquisition system of the speed reducer, the equipment maintenance record database, the equipment operation log, etc., and collect the operation data of the speed reducer under different working conditions and at different times; the operation data includes physical parameters such as rotational speed, torque, temperature, vibration frequency, vibration amplitude, etc., and the operation status of the equipment, such as normal operation, fault status, etc.

[0045] Clean the collected historical data to remove the noise data, outliers and missing values. For example, for the vibration amplitude data, if a certain data point significantly deviates from the normal range, it may be caused by sensor failure or data transmission error, and it needs to be removed or corrected.

[0046] After cleaning, perform feature extraction operations; in feature selection, according to the operating principle and fault characteristics of the speed reducer, select the features related to fault diagnosis and prediction from the cleaned historical data. For example, for the gear fault of the speed reducer, the spectral characteristics of the vibration signal have important diagnostic value, such as meshing frequency, sideband, etc., so these features can be selected as the input of the model. At the same time, deeply mine the historical data of the speed reducer, analyze the change trend of the operating parameters of the speed reducer over time. For example, analyze the change law of parameters such as vibration amplitude and temperature over a period of time to judge whether there are abnormal fluctuations or trend changes. Study the correlation relationship between the operating parameters of the speed reducer, and find out the feature combinations closely related to the occurrence of faults. For example, analyze the influence of the changes in rotational speed and torque on the vibration signal and their correlation with gear faults.

[0047] Next, further process and transform the selected features. Use Fourier transform to convert the vibration signal in the time domain into a frequency domain signal and extract spectral features; use wavelet transform to perform multi-scale analysis on the signal to obtain feature information in different frequency bands; improve the performance of the model through processing and transformation.

[0048] According to the characteristics of the problem and the features of the data, select appropriate diagnostic models, including machine learning models and deep learning models; machine learning models such as support vector machines, decision trees, random forests, etc., and deep learning models such as convolutional neural networks, recurrent neural networks, etc.

[0049] After selecting the appropriate diagnostic model, perform model training. Use the extracted features and corresponding fault labels to train the selected model. During the training process, by adjusting the parameters of the model, enable the model to accurately learn the fault features and patterns in the historical data of the reducer. Use the operating parameters in the historical data and the corresponding fault occurrence time to train the selected prediction model, so that the model can learn the relationship between the change trend of the reducer operating parameters and the occurrence of faults.

[0050] Use methods such as cross-validation and holdout method to evaluate the trained model, calculate indicators such as the accuracy, recall rate, and F1 value of the model, and determine whether the performance of the model meets the requirements. If the model performance is poor, it is necessary to adjust and optimize the model, such as adjusting the parameters of the model, increasing the feature dimension, replacing the model, etc.

[0051] Among them, in the step of obtaining and outputting the reducer fault diagnosis data according to the reducer data mapping table, define the mapping relationship between the reducer operating data and the fault type according to the structure and operating principle of the reducer. For example, when the vibration frequency of the reducer is within a certain specific range, it may correspond to a gear wear fault; when the temperature exceeds a certain threshold, it may correspond to a poor lubrication fault. Organize the defined mapping relationship into the form of a data mapping table for convenient subsequent query and use. The data mapping table can be stored in the form of a database table, Excel table, etc. Real-time obtain the preprocessed sensor and image multi-modal data, and according to the sensor and image multi-modal data, the diagnostic model predicts the operating state of the reducer and outputs possible fault types. According to the fault type output by the model, query information such as the corresponding fault description, possible causes, and solutions in the data mapping table.

[0052] Among them, in the output of the fault diagnosis result, adopt the method of visual display and report generation: Visual display, display the fault diagnosis result to the user in a visual way, such as displaying information such as the operating state, fault type, and fault location of the reducer through a graphical interface.

[0053] Report generation: Generate a detailed fault diagnosis report, including the time and location of the fault, fault type, cause of the fault, solution, etc., and send the report to relevant personnel.

[0054] In the step of adopting the diagnostic model, obtaining the fault prediction data of the speed reducer according to the speed reducer data mapping table, and outputting: Input the real-time collected operation data of the speed reducer into the trained prediction model, predict the future operation state of the speed reducer according to the learning results of historical data, and output the possible fault types and occurrence times within a certain period in the future. According to the fault types output by the prediction model, query the corresponding fault descriptions, possible causes, preventive measures and other information in the data mapping table.

[0055] In the output of the fault prediction result, adopt the method of warning prompt and report generation: Warning prompt: When it is predicted that the speed reducer may have a fault, send a warning prompt to relevant personnel in time to remind them to take corresponding preventive measures. The warning prompt can be sent by text message, email, system message, etc.

[0056] Report generation: Generate a detailed fault prediction report, including the predicted fault type, occurrence time, possible cause, preventive measure, etc., and send the report to relevant personnel so that relevant personnel can formulate a reasonable maintenance plan.

[0057] In this step, through the above steps, the fault diagnosis and prediction of the speed reducer can be realized, providing strong support for the maintenance and management of the equipment.

[0058] S300: Obtain the associated data of the speed reducer and dynamically output the working mode of the strip mill.

[0059] In this embodiment, obtain the associated data of the speed reducer and dynamically output the working mode of the strip mill. The specific process is as follows: S301: According to the cooperation data of multiple speed reducers in the steel mill, associate each speed reducer in single and multiple steel mills; S302: Divide the task types of the steel mill and define the priorities of the task types; S303: Divide the operation area, obtain the emergency task type according to the fault diagnosis and prediction data of the speed reducer, query the current task instructions of multiple steel mills in the operation area, compare the priorities of the current task type and the emergency task type, and output the next operation instruction for the steel mill.

[0060] In the above process, each reducer in a single or multiple steel rolling mills is associated according to the collaboration data of multiple reducers in the steel rolling mill; the number of reducers required for different production situations is defined, and the reducer association data is formulated; for example, in the case of a single reducer configuration, low-capacity rolling production can be carried out; for two reducer configurations, driving the upper and lower rollers, high-precision rolling production can be carried out; for multiple reducer parallel drive configurations, thick plate rolling production can be carried out.

[0061] Divide the task types of the steel rolling mill and define the priorities of the task types; the task types can include emergency tasks, timing tasks, and temporary tasks; among them, the priority of emergency tasks is level one, the priority of timing tasks is level two, and the priority of temporary tasks is level three.

[0062] Divide the operation areas in the production process, obtain the sudden task types according to the reducer fault diagnosis and prediction data, query the current task instructions of multiple steel rolling mills in the operation area, compare the priorities of the current task type and the sudden task type, and output the next operation instructions for the steel rolling mill.

[0063] Among them, in the comparison of the priorities of the current task type and the sudden task type: When the priority of the current task type is higher than or equal to the priority of the sudden task type, output the instruction for the steel rolling mill to continue the current task; When the priority of the current task type is lower than the priority of the sudden task type, output the instruction for the steel rolling mill to terminate the current task and package the current task data.

[0064] In this step, according to the reducer fault diagnosis and prediction data, the working modes of single and multiple strip steel rolling mills in the operation area are dynamically output, where the working modes can include low-capacity rolling production mode, high-precision rolling production mode, thick plate rolling production mode, etc., so as to reduce the occurrence of unplanned shutdowns and optimize the strip steel production rhythm.

[0065] In the present invention, first, a unique identifier is assigned to the reducer of the strip steel rolling mill, multi-modal data of sensors and images is obtained, and a reducer data mapping table is established; then a diagnostic model is constructed, the reducer data mapping table is obtained, and the reducer fault diagnosis and prediction data is output; finally, the reducer association data is obtained, and the working mode of the strip steel rolling mill is dynamically output; by dynamically adjusting the working mode of the strip steel rolling mill according to the fault diagnosis and prediction of the reducer of the strip steel rolling mill, the occurrence of unplanned shutdowns is reduced, and the strip steel production rhythm is optimized.

[0066] Corresponding to the embodiments of the multi-modal fusion-based intelligent fault diagnosis and prediction method for the reducer of the strip steel rolling mill described above, this application also provides embodiments of a multi-modal fusion-based intelligent fault diagnosis and prediction system for the reducer of the strip steel rolling mill.

[0067] Figure 5It is a block diagram of an intelligent fault diagnosis and prediction system for a strip mill reducer with multimodal fusion shown according to an exemplary embodiment. Refer to Figure 5 , the system may include: a data mapping table establishment module 401, a diagnosis and prediction module 402, and a mode dynamic output module 403; wherein: The data mapping table establishment module 401 is used to assign a unique identifier to the strip mill reducer, obtain multi-modal data of sensors and images, and establish a reducer data mapping table; The diagnosis and prediction module 402 is used to construct a diagnosis model, obtain the reducer data mapping table, and output reducer fault diagnosis and prediction data; The mode dynamic output module 403 is used to obtain reducer-related data and dynamically output the working mode of the strip mill.

[0068] In this embodiment, the data mapping table establishment module 401 assigns a unique identifier to the strip mill reducer, obtains multi-modal data of sensors and images, and establishes a reducer data mapping table; the diagnosis and prediction module 402 constructs a diagnosis model, obtains the reducer data mapping table, and outputs reducer fault diagnosis and prediction data; the mode dynamic output module 403 obtains reducer-related data and dynamically outputs the working mode of the strip mill; by diagnosing and predicting the faults of the strip mill reducer, the working mode of the strip mill is dynamically adjusted, the occurrence of unplanned shutdowns is reduced, and the strip production rhythm is optimized.

[0069] Regarding the system in the above embodiment, the specific manners in which each module performs operations have been described in detail in the embodiment of the method related thereto, and will not be elaborated herein.

[0070] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present application. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0071] Correspondingly, the present application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the multimodal fusion intelligent fault diagnosis and prediction method for the strip mill reducer as described above. As Figure 6As shown, it is a hardware structure diagram of any device with data processing capabilities where the intelligent fault diagnosis and prediction system for the reducer of a strip mill with multi-modal fusion provided by an embodiment of the present invention is located. Besides Figure 6 the shown processor, memory, and network interface, any device with data processing capabilities where the device in the embodiment is located usually may further include other hardware according to the actual functions of the any device with data processing capabilities, which will not be elaborated herein.

[0072] Correspondingly, the present application also provides a computer-readable storage medium, on which computer instructions are stored. When the instructions are executed by a processor, the intelligent fault diagnosis and prediction method for the reducer of a strip mill with multi-modal fusion as described above is implemented. The computer-readable storage medium may be an internal storage unit of any device with data processing capabilities described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium may also be an external storage device, such as a plug-in hard disk, a Smart Media Card (SMC), an SD card, a FlashCard, etc., equipped on the device. Further, the computer-readable storage medium may also include both an internal storage unit of any device with data processing capabilities and an external storage device. The computer-readable storage medium is used to store the computer program and other programs and data required by the any device with data processing capabilities, and may also be used to temporarily store data that has been output or will be output.

[0073] Those skilled in the art will readily think of other implementation schemes of the present application after considering the specification and practicing the content disclosed herein. The present application aims to cover any variations, uses, or adaptive changes of the present application, and these variations, uses, or adaptive changes follow the general principles of the present application and include common general knowledge or conventional technical means in the technical field not disclosed in the present application.

[0074] It should be understood that the present application is not limited to the exact structure already described and shown in the drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method, characterized in that: The steps include: Assign a unique identifier to the strip mill reducer, obtain sensor and image multimodal data, and establish a reducer data mapping table; Build a diagnostic model, obtain the reducer data mapping table, and output the reducer fault diagnosis and prediction data; Obtain reducer-related data and dynamically output the strip mill working mode.

2. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 1 is characterized in that: In the steps of assigning a unique identifier to a strip mill reducer, acquiring sensor and image multimodal data, and establishing a reducer data mapping table: Define coding rules and assign unique identifiers to reducers based on information from the strip mill; A variety of sensing devices and image acquisition devices are arranged at the reducer to respectively acquire the temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the reducer; Create a reducer data mapping table using the unique identifier as the primary key.

3. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 2, characterized in that: In the step of arranging a plurality of sensing devices and image acquisition devices at the reducer to respectively acquire temperature data, pressure data, vibration amplitude data, lubricating oil level data, and lubricating oil image data of the reducer: The data obtained by the sensor device and the image acquisition device are given a timestamp and a unique identifier of the reducer.

4. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 3 is characterized in that: Before the step of creating a reducer data mapping table with a unique identifier as the primary key: Use the unique identifier as the main search term to query the data of the unique identifier and perform data aggregation.

5. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 1, characterized in that: In the steps of building a diagnostic model, obtaining a reducer data mapping table, and outputting reducer fault diagnosis and prediction data: Obtain historical data of the reducer and use the historical data of the reducer to build a diagnostic model; Using the diagnostic model, according to the reducer data mapping table, the reducer fault diagnosis data is obtained and output; The diagnostic model is used to obtain the reducer fault prediction data according to the reducer data mapping table and output it.

6. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 1, characterized in that: In the step of obtaining reducer-related data and dynamically outputting the working mode of the strip mill: According to the collaborative data of multiple reducers in the steel rolling mill, each reducer in a single or multiple steel rolling mills is associated; Classify steel mill task types and define task type priorities; Divide the operation area, obtain the sudden task type according to the reducer fault diagnosis and prediction data, query the current task instructions of multiple steel rolling mills in the operation area, compare the priorities of the current task type and the sudden task type, and output the next operation instruction for the steel rolling mill.

7. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 6, characterized in that: In the steps of classifying steel mill task types and defining the priorities of task types: Task types include emergency tasks, time-series tasks, and temporary tasks; The priority of urgent tasks is level one, the priority of sequential tasks is level two, and the priority of temporary tasks is level three.

8. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 6, characterized in that: In the step of comparing the priorities of the current task type and the sudden task type and outputting the next operation instruction for the steel rolling mill: When the priority of the current task type is higher than and equal to the priority of the emergency task type, the output steel rolling mill continues the current task instruction.

9. The multi-modal fusion strip mill reducer fault intelligent diagnosis and prediction method according to claim 6, characterized in that: In the step of comparing the priorities of the current task type and the sudden task type and outputting the next operation instruction for the steel rolling mill: When the priority of the current task type is lower than the priority of the burst task type, the output steel rolling mill terminates the current task instruction and packages the current task data.

Images